Derivative Linear Sweep Voltammetry and Discrete Wavelet Transform for the Simultaneous Determination of Codeine and Thebaine by Artificial Neural Networks

摘要

The main problem that limits the application of Linear Sweep Voltammetry (LSV) in simultaneous determinations is the presence of background interference, which affects the accurate multicomponent determinations due to the overlapping of the voltammetric signals. This study represents the chemo-metrics processing of a series of the first-derivative LSVs, recorded on a glassy carbon electrode modified with Multi-Walled Carbon Nanotubes (MWCNTs) and Poly L-methionine, using Discrete Wavelet Transform (DWT) to compress the resulting data and feedforward back-propagation Artificial neural networks (ANNs) for the simultaneous determination of codeine and thebaine, as two opium alkaloids. Linear univariate calibration curves using first-derivative voltammograms were obtained in the ranges 2-1000 μM for codeine (at Ep = 814 mV) and 5-500 μM for thebaine (at Ep = 741 mV), respectively. Besides, the excellent regression parameters for the comparison graphs of ANNs and satisfactory recoveries in real serum samples in the ranges 98-102% for codeine and 90-107% for thebaine indicate the applicability of the proposed method for routine simultaneous analyses.